The generation of visually guided movement takes place in a dense network comprised of many interconnected cortical regions. To study this process, we have developed a task based on a motor illusion that dissociates visualization of drawing from the movement used to produce it. Multielectrode recording techniques have been proven and will be used to record single-unit activity from populations of neurons in three nodes of the parietal-frontal portion of the network. A number of advanced analyses will be used to identify groups of cells with activity related to different aspects of this behavioral task. The conventional population vector algorithm will be applied to large populations of cells and a new, more sensitive;particle filter will be used on small groups. Both of these population algorithms wilt be used to extract neural representations of the arm trajectory (drawn shape). With these neural trajectories and the illusion task, it will be possible to determine whether a population is related to the visualized or executed movement, a determination that we have shown (in preliminary data) can be made for at least two of the nodes to be studied. These techniques will allow us to measure timing of the representations relative to the movement and thereby ascertain which groups are acting as feed forward and feedback elements of the control process. This will help build a schematic of the overall process used in visuomotor behavior.